Singularly Stupid | Hacker News Reader

Singularly Stupid

blogs.law.harvard.edu

54 pointsbdfh4241 comments

"Philip Greenspun Writes About the Singularity," or, "How to Do It Wrong!"

What would you call people who pay $25,000 for a nine-week course of study with a collection of Silicon Valley optimists? “Singularly Stupid”?

Name calling. This can in no way contribute to any discussion where understanding is the goal. Arguably, it follows that the rest of the essay is not intended to help people's understanding.

The idea of the singularity is that technology, especially in the form of artifically intelligent robots, will solve all of our problems and technological advance will speed up exponentially starting roughly around the year 2030.

The point is that the rate of advances in nearly all fields is exponential. Doubly exponential, in fact. That is, the exponent is growing at an exponential rate. And it's not that all of our problems will be solved, it's that the advances will be so dramatic that rules derived from how things are now--this side of the singularity--are useless for understanding what happens after the singularity.

So far technology innovation hasn’t outstripped Malthusian human population growth. We can grow more food more efficiently, but the number of human mouths to feed has grown just about as fast, so that we struggle to feed everyone. A lot of what we’ve done over the past few hundred years has come at the cost of using up the Earth, e.g., clearing forest for farmland or digging up coal and oil and lighting it on fire, taking all of the Cod out of the North Atlantic.

One of the most blatant straw men I've seen in a long time. First, technological innovation hasn't outstripped Malthusian population growth because technological innovation is the cause of it! Frankly, there are a lot of sad things about having 6 billion people on earth, but I think that about 6 billion of them would vote against programs that say, reduced the food supply.

Second, this ignores the amount of stuff we are able to get out of a unit of matter taken from the environment, and the amount of pollution produced by a unit measure of production, and the amount of dispersal for a given amount of pollution. In short, it categorically ignores any good that can have come from technology.

Far from freeing us from cleaning the house, Artificial Intelligence thus far has failed to live up to promises made by professors seeking research funding in 1960 (that reminds me I need to do laundry!).

It's probable that this fallacy has a proper name, but I've always called it Ascription of Intent. And I'm talking about the first implicit fallacy, which is an assertion that professors lied about AI for the sole reason of obtaining funding. The other fallacy is an ad hominem attack, namely that because professors were (allegedly) lying to get funding, AI is [useless|stupid|malevolent|a bad thing].

Given the track record of tech as a mixed blessing and as a slower agent for change than predicted, do young people need to prepare for 2030? Can they prepare by listening to Ray Kurzweil, or anyone else born in 1948? Should they fork over $25,000 for nine weeks or simply watch old Jetsons episodes?

It's coming thick and fast, now. Technology is a slower agent of change than predicted? Really? Who's predictions? I remember a prediction about computers someday being smaller than a rhinoceros, and another about how every town will have one telephone. Even 20 years ago, did anyone imagine that a global computer network would be so pervasive that even logically fallacious essays would be conveyed instantaneously the world over?

Furthermore, his point is specifically and directly addressed by Kurzweil et al. Actually, not even 'et al.' In 'The Singularity Is Near,' Kurzweil talks about how even scientists--especially scientists--underestimate the pace of progress because of their proximity to the drudgery of their own various specialties. He talks about how the aggregate adds up more quickly than humans are able to directly sense, and that it takes analysis to get a true picture of the rate of progress. I think we need to insist that Greenspun at least obey the principle of charity and put the best representation of his opponent's argument forward, rather than the worst. (And not just for the sake of charity, but because that is how collective knowledge--rather than, for example, an author's personal agenda--would be advanced.)

And what does anyone's birth date have to do with it? He may as well have comment on the race, creed, national origin, sexual orientation, or eye color of Singularity advocates.

Maybe I will kick off the comments section with a realistic tech innovation that would change the world in a positive way. My pick: A better battery (cheaper, lighter, higher power density). That would enable the use of renewable energy in every kind of portable application, e.g., cars and airplanes, and also make it much more practical to use wind and solar generation.

Has he not heard of Lithium-Ion batteries? He's not heard of the Tesla Roadster? Is he asserting that battery technology is the only factor limiting the adoption of renewable energy? Power generation density isn't an issue? High capital costs aren't an issue? Regulatory hurdles aren't an issue? It's batteries that are going to save us?

The "[s]pecial offer' at the end doesn't even rise to the level where comment is necessary.

Addendum:

There has been a lot of hand-wringing recently over the direction that HN has been taking. That this article has made the front page should be taken as food for thought in this regard. The content of this article didn't even rise to the level of 'commentary' or 'argument.' It's easy to get off track into name-calling, non-sequiters, and simple wrong-headedness; this applies to comments as well as story submissions. I'm not here arguing whether this article should have been front-paged. I am arguing that, if you care about HN, one of the things you can do is make sure that your comments do not resemble this article in vacuity, stupidity, needless over-emotionalism, or fallaciousness.

Just a nit: Greenspun is wrong about food production technology. Food scarcity is no longer a problem: we have enough to feed everyone in the world, and make some of them fat.

Distribution is an issue in some places, but production is a solved problem.

Can food production technology stay apace with a couple billion Chinese and Indians starting to eat like Westerners (lots of beef, which requires tremendous amounts of grain and water to produce)?

That's not a scarcity problem but a scarcity of luxury problem.

Scarcity of luxury is by definition a subset of scarcity, so it is a scarcity problem.

Sure, we could all live on beans on rice, but people don't do that unless they have to.

But Greenspun's claim is that technology has done nothing to ease Malthusian limits, and he's wronger now than he's ever been. At the time Greenspun wandered into this fray, the concern was over human starvation. The situation really is much, much better, to a degree that is hard to fully appreciate.

Don't worry, they won't eat more than what is produced.

Are you sure? Hint: agflation.

Production is solved for now, but our solution (at least in the US, as I understand it) depends on abundant supplies of fossil fuel and fossil water, which are not going to be around forever.

So, there is plenty of work to do to see that this problem stays solved.

Abundant supplies of fossil fuel's are in use because they are cheep. Zero technological advances are required to provide a stable food supply for 20 billion people. It's actualy difficut to find the maximum stable food supply with current technology over the long term.

For example: "there are 1.3 billion cattle in the world today"

Now some of them are a net food source because they feed in areas that humans can't. But, effectively feeding cows costs the world well over 1 billion humans worth of food.

Topsoil does not actually get destroyed. Rivers just deposit it as silt which can be recovered.

etc.

Topsoil does not actually get destroyed. Rivers just deposit it as silt which can be recovered.

Interesting. Silt also piles up behind dams. http://www.google.com/search?q=irn+dams+silting Perhaps dam silt could be even-more easily recovered.

feeding cows costs the world well over 1 billion humans worth of food.

RBST can reduce that impact. http://www.google.com/search?q=rbst+efficiency

rbST reduces carbon footprint of dairies

The present study demonstrates that use of rbST markedly improves the efficiency of milk production and mitigates environmental parameters, ...

fossil water is not an issue - aquifers are constantly replenished with rainwater (and manual recharging via pumping treated water back into the aquifer...a debatable idea). At least that's the case in the mid- and north-west. Coastal and dryish states may vary. I seem to remember seeing somewhere that Los Angeles was putting an evaporative water purification plant in, which would give the aquifers in that area a rest.

fossil fuel, however, is still an issue.

It's not exactly a nit. Greenspun's worldview has been proven wrong for two centuries.

Also, Greenspun seems not to be aware that wealth is correlated with improved environmental conditions after a certain point.

If this short post had something to offer besides unnuanced hostility, I might be more inclined to forgive its errors.

Greenspun is also wrong about what singularity is. Afaik its the point where human as intelligent agents build a better intelligent agent than ourselves. That new agent will do the same ,leading to agents of increasing intelligence. There isnt a date fixed on that event (as he says) and the debate is still whether this is posible or not. You have an interesting essay on that at http://www-rohan.sdsu.edu/faculty/vinge/misc/WER2.html

Yeah, I thought it wasn't in 2030, but has been defined as $now + 20 years.

Apologies if this is off topic, but I've learned to associate a harvard.edu URL in the HN topic with a Philip Greenspun post. I try to avoid his writing, so if other people at Harvard have interesting things to say, I'm not going to find it through Hacker News.

The same thing happens with subdomain blogs on github - I can't tell which blog the post is really coming from. The posts are rarely from GitHub, the company.

Is the reason the domain is used rather than the entire URL for aesthetic reasons? The short domain looks better. However, I have an internal filter of what I do and don't want to click on and seeing the entire URL helps.

"So far technology innovation hasn’t outstripped Malthusian human population growth."

Er, yes, it has. There's more food/person on the planet now than when Malthus wrote.

I have to disagree with Dr. Greenspun here. Despite the Singularity topic's lack of immediate value some people are still be interested in studying with some of the most famous futurists in the world. Many people (myself included) enjoy learning for the challenge of thinking in new ways and this program at least promises that opportunity.

Yeah - I think though that the $25,000 pricetag is outlandish. It should be free. The benefit to society will be realized in the connections that result from brining together these 'minds' rather than from whatever topic they will or won't learn while there.

I think its a good step though - to actively foster innovation rather than assume academia (which I guess this is, though non-standard - which is good) business or war will generate it.

Did you miss the part about the $25,000? Best as I can tell from the TED website, that's the admission to twelve TED conferences. Or two to three years of state university tuition (depending on your home state). Or a very very large book collection. Or one hell of a lot of plane tickets to see the lectures that every "famous futurist" gives for free, or for a relatively modest conference fee, multiple times per year. (If they didn't give public talks all the time these people wouldn't be famous futurists! And I assure you that the more obscure futurists, toiling away in academia, are almost certainly available to chat with you at a personal one-on-one lunch for a lot less than $25,000, or even $2000.)

But, you know, one of the side effects of population growth is that there's now more than one sucker born every minute.

UPDATE: Someone else in the thread says TED costs $6000 for four days. So, only four TED conferences. Or something. If you have to ask you probably can't afford it. ;)

I remember hearing an interview on NPR where a stamp collector calculated the cost of a stamp at ~30,000$ per time he actually looked at it in his vault. For you or me that might seem stupid, but once you have more income than you can reasonably spend in your lifetime numbers below a million stop being meaningful.

Or as a factory worker in china might say you spent 7.50$ on a cup of coffee?

Whether or not the program is worth $25,000 is up to the market to decide, not just you. Basing the value of a program by arbitrarily declaring "that's just too much money for me!" does not address whether or not the program ultimately offers enough value to the actual students. That question will be resolved in time.

If I walk by a roulette wheel and I see someone put $10,000 on black, I deduce that the person who laid down that money is an idiot. I realize that one can make sophisticated arguments about "risk preference" and the idea that the market sets a price for everything and so on, but at a certain point you just have to say that the emperor has no clothes.

Or two rounds of YC funding? :-D

Perhaps, in exchange for the $25K, you will learn how to convince our future robot overlords to upload your brain into one of the good mainframes.

I saw an article in IEEE magazine about the singularity and it seemed more about people talking about the implications of technology rather than actually working to build technology. The singularians seem to be a group of very capable technologists - why don't they allocate their time and resources to improving technologies that could make the singularity happen instead of talking about it? Is hyping the singularity a meta-strategy to attract the resources and effort of others?

Charging 25k is a strong sign of cultish scam behavior. Go figure.

They are. There are many singularitarians working on concrete things in AGI/AI, life enhancements, cryogenics, biology... but of course, you'll hear more about the speakers and futurists, since that's what the news will go after.

I'm tired of hearing about a better battery. Yes it would be useful and lucrative, but it's a very hard problem that has been worked on by lots of people for a long time.

I think that's the point of this example. If technology is improving at an exponentially accelerating rate, how come we have all these technologies sitting around that aren't improving at all, despite lots and lots of effort?

Sterling makes a similar point in his hilarious deconstruction of the Singularity, which IMHO is required listening for students of this topic:

http://media.longnow.org/seminars/salt-0200406-sterling/salt...

He points out that Singularitarians go on and on about the spectacular technology improvements in data networks, but manage to ignore all the other networks -- plumbing, railroads, highways -- that are much more important to our lives but which haven't changed technologically since the 1950s or before.

I think Kurzweil's viewpoint is simply to point out the exponential growth in information technology, and how that may affect humanity, rather than to incorporate and understand the effects of ALL technology.

On the radio recently, I heard that Spain is improving its rail infrastructure so that all cities will be within four hours of train travel from Madrid in the next five years - that's 2014.

That made me realize that if Kurzweil's extrapolations towards a singularity are roughly right, the imminent-singularity± world will look a lot like the world looks now. Roads will be much the same, people wont have their houses rebuilt, office buildings will be similar. It will be all the technology inside that's different.

If you've read Vernor Vinge's latest novel, Rainbow's End, that feels like a fairly convincing future. Mostly augmented reality, fewer user serviceable parts, and improved medical technology.

±imminent-singularity implies a sharp switch. That's not how I think of it - it's more a series of concentric circles - singularity event horizons, future-shock waves emanating back into the past - that we are already in, and have always been in, where each one marks a point where enough quantitative change has happened to make a qualitative change that would be difficult to predict from before the change. The imminent-singularity refers to the fact that they are getting closer together, and as more of them happen in less time, the cumulative effects are that much stronger and faster, leading to a techno-sonic-boom.

The key changes to hurry along "the singularity" are changes to hurry along new and better kinds of thought. As Alan Kay(?) said, computers are tools for thinking, so I think it's right to look at computing/IT for the changes with the most impact.

Significant AI work will, I think, be software based, or biological designed with software. Significant nanotechnology design will be done with lots of aid from computers. Significant collaborative work will be done with the help of future telecoms infrastructure. Significant sharing and spreading of ideas will be telecoms related.

Plumbing, railroads, highways ... A new kind of collaborative shower? Trains that run even if there are leaves on the line? What kind of technological change do you think would make major differences? I think something along the lines of personal communications tech would make the most difference, e.g.:

- An open API for all train and bus timetables and up to date changes. Imagine being in an unfamiliar city, opening Google maps. It finds your location, you click where you want to go, it finds a bus/underground route for you, next few buses, option to buy a ticket.

- Local area networking such that you walk into an airport, your phone knows which flight you need so arrows appear on it and show you where to go in the airport. Over here, down there, to this checkin, then to that lounge.

The technological changes that I think will make a difference are in people and bureaucracy and communication, not in roads and rail themselves.

Why not simply take a taxi cab?

Because that wouldn't illustrate my point to the previous poster that improvements for road/rail will be to do with people and information, not so much cars or trains directly. :)

Why not use public transport?

- Because you don't know where it is. Enter technology to point you to a bus/train station. Enter GPS on your phone to guide you to it.

- Because you don't know which route / bus / train to take. Enter an open transit information system so your phone can tell you, on the fly.

- Because you don't know if this is the right bus/train to get on. Enter QR codes on the side of the bus so your phone can confirm it for you.

- Because you don't know which stop to get off at. Enter GPS on your phone to tell you.

- Because you don't know which ticket to get or if you have the right change. Enter your mobile and your phonebill or other integrated low-hassle small-quantity payment service.

I'm pretty sure the $25k is for those who Google and/or other sponsors don't give a scholarship to. Its a nice way of saying if we haven't invited you, you can come; but BYOB (bring your own bank)

I feel like the first part of that was a straw man argument.

But the conclusion had one good point.

The problem I have with singularity folks is a matter of putting the horse before the cart.

Sure it's nice to dream about the world after the singularity, but isn't your time better spent taking one step after another towards that future?

And doesn't that require taking the first step, which would be something improving today's technology?

Most singulatarians I speak with understand that thinking about what life will be like after the singularity is fundamentally pointless. One of the definitions of the singularity is the point beyond which extrapolation of current trends becomes impossible. Most of my discussions with singulatarians focus on the nature of technological innovation and which of several possibilities for augmented or artificial intelligence might come first.

Vis-à-vis the nature of technological innovation:

I have this crazy notion that more is more.

In other words, the reason the valley is first in .com and other tech innovation is because it has the most resources, in terms of people, funding, etc, working on that.

So the best way to speed up innovation is to do everything you can to get more people working on innovation.

That could include lobbying, or educating yourself, or teaching, etc.

Sure, $25,000 is a lot for 9 weeks. But TED, where SU was officially announced, costs $6,000 for 4 days.

Also, maybe they'll use exponential progress to halve their fees every year. It'll cost under $1000 in 5 years, and under $1 in 15!

Harvard tuition in 1993: $23,514

Harvard tuition in 2008: $45,620

Harvard tuition in 2023: Potentially, under $1. But I'm not betting on it.

sources:

http://vpf-web.harvard.edu/budget/factbook/00-01/page23.html

http://www.fao.fas.harvard.edu/cost.htm

I think a day of brilliant optimism would be a real treat. Optimism without brilliance may be a bit much. But brilliance especially with optimism added sounds great.

Reply on news.ycombinator.com